I have a following problem. Given a dataset with var X. I like to use the obs in this dataset as a base to come up with a distribution, (not necessarily any known distribtution) and then begin to randomly simulate obs from this distribution. I was wondering if there is a proc in SAS that does that, or will help me doing this. Hopefully, I like to specify if the distribution is discrete or continuous with given ranges. I have something like that in mind for discrete case.

>From dataset T, generate a table with X Freq Probablity 1 300 0.3 2 200 0.2 4 500 0.6.

If we assume that 1,2,4 are the only values that are assumed by X, then we can randomly simulate this X from the distribution which is calculated based on freq.

I can see it is more difficult to deal with a continuous case, there are some methods in the literature of simulation. But, I was hoping that SAS has a procedure will do the following directly,

x_obs=Call SASPROC(dataset) , there dataset is the input dataset from which a distribution is determined and SASPROC will return a random var from this distribution.

Your input and info is greatly appreciated.

Thanks,

Hongjie Wang

ZS Associates Princeton, NJ 08536